Talk of genetics and vice versa

Abstract

Does our ability to talk lie in our genes? The suspicion is bolstered by the discovery of a gene that might affect how the brain circuitry needed for speech and language develops.

Full text

“M
an has an instinctive tendency
to speak, as we see in the babble
of our young children,” wrote
Charles Darwin
1
in 1871, “while no child has
an instinctive tendency to bake, brew, or
write.” Darwin’s observation has just been
supported in a way he could not have dreamed
of, with the discovery by Lai and colleagues
2
(page 519 of this issue) of a gene that is mut-
ated in a disorder of speech and language.
The possibility that human language abil-
ity has genetic roots was raised about forty
years ago by the linguist Noam Chomsky
3
and
the neurologist Eric Lenneberg
4
. Chomsky
noted that language is universal, complex
and rapidly acquired by children without
explicit instruction. Lenneberg pointed out
that a small number of children fail to display
this talent and that such deficits sometimes
run in families. Deficits of this kind are
now called ‘specific language impairment’,
an umbrella term for language disorders that
cannot be attributed to retardation, autism,
deafness or other general causes. Specific
language impairment not only runs in fami-
lies but is more concordant in identical than
in fraternal twins, suggesting that it has a heri-
table component
5
. But the inheritance pat-
terns are usually complex, and until recently
little could be said about its genetic basis.
Then, in 1990, investigators described the
‘KEs’ — a large family, of several generations,
in which half the members suffer from a
speech and language disorder
6
. This disorder
is distributed within the family in a manner
that suggests it is caused by a dominant gene,
or a set of linked genes, on an autosomal
(non-sex) chromosome. The press referred
to it as a ‘grammar gene’ (Fig. 1), while scep-
tics suggested that it merely lowers intelli-
gence or makes speech unintelligible, or
even that the disorder is nothing more than
an artefact of a working-class dialect.
Extensive testing by psycholinguists,
including Faraneh Vargha-Khadem, one of
the authors of the paper in this issue
2
,
suggested that the disorder is more complex
than either of these extremes
7,8
. Affected
family members do tend to score below
average in intelligence tests (perhaps because
verbal coding helps performance in a variety
of tasks). But the language impairment can-
not be a simple consequence of low intelli-
gence, because some of the affected members
score in the normal range, and some score
more highly than their unaffected relatives.
And although the affected members have
problems in articulating speech sounds
(especially as children) and in controlled
movements of the mouth and tongue (such
as sticking out their tongue, or blowing on
command), their language disorder cannot
be reduced to a problem with motor control.
They also have trouble identifying basic
speech sounds, understanding sentences,
judging grammaticality, and other language
skills. For example, as adults they stumble
at a task involving nonsense words that most
four-year-olds pass with ease: completing
sequences such as ‘Every day I plam; yester-
day I _____’
9
.
In 1998 several of the authors of today’s
paper linked the disorder to a small segment
of chromosome 7, which they labelled
SPCH1 (ref. 10). Now, thanks to the discov-
ery of an unrelated person known as CS, who
has both a similar speech deficit to the KEs
and a chromosomal translocation affecting
the SPCH1 segment, Lai et al.
1
have nar-
rowed the disorder down to a specific gene,
FOXP2. In CS, this gene is disrupted by the
translocation. In all the affected members
of the KE family examined, but in none of
the unaffected members, and in none of 364
chromosomes from unrelated, unaffected
people, a single guanine nucleotide is
replaced by an adenine. (The perfect contin-
gency is in striking contrast to the now-you-
see-it, now-you-don’t correlations found in
the first generation of searches for genes
affected in behavioural disorders.) The
authors propose that the nucleotide replace-
ment results in substitution of the amino
acid histidine for an arginine in one struc-
ture — the ‘forkhead’ domain — in the
gene’s protein product, presumably altering
the protein’s function.
Lai et al. present hints that FOXP2 may
have a causal role in the development of the
normal brain circuitry that underlies lan-
guage and speech, rather than merely dis-
rupting that circuitry when mutated. FOXP2
belongs to a family of genes that encode
transcription factors (proteins that trigger
the copying of genes into messenger RNAs),
many of which have important roles in
embryonic development. One of the defin-
ing features of proteins in this family is the
forkhead domain, which contacts a target
region in DNA, and it is this domain that is
affected by the mutation in FOXP2. FOXP2
appears to be strongly expressed in fetal
brain tissue (among other places), and its
homologue is expressed in the developing
cerebral cortex of mouse embryos. In both CS
and the affected members of the KE family,
only one copy of FOXP2 is disrupted. So Lai
et al. suggest that, at a critical point in fetal
brain development, affected individuals
have only half the normal amount of func-
tioning transcription factor, which is not
enough to control some aspect of early brain
development.
Whatever the exact function of the gene
turns out to be, the new work
2
has many
implications. As a smoking gun for a genetic
cause of one kind of language disorder, the
discovery motivates the search for genetic
causes of cognitive and learning disorders
more generally, relieving the presumption
of guilt from mothers (who are often still
blamed for everything that goes wrong
with their children). It also shows that just
because a cognitive disorder has a genetic
cause, it is not necessarily untreatable. The
affected KE adults learned to compensate
for their difficulty in generating complex
linguistic forms by memorizing the forms
NATURE
|
VOL 413
|
4 OCTOBER 2001
|
www.nature.com 465
news and views
Talk of genetics and vice versa
Steven Pinker
Figure 1Genes and speech: a cartoonist’s view of
a ‘language gene’. The identity of a gene affecting
speech and language has been pinned down by
Lai et al.
2
, writing in this issue. (Reprinted with
special permission, North America Syndicate.)
Does our ability to talk lie in our genes? The suspicion is bolstered by the
discovery of a gene that might affect how the brain circuitry needed for
speech and language develops.
© 2001 Macmillan Magazines Ltd

whole and by consciously applying rules they
had been taught in language therapy
11
. These
and other strategies allow them to converse
competently, although this has made life dif-
ficult for psycholinguists trying to work out
the underlying disorder from the behaviour
of affected adults.
If FOXP2 really does prove necessary for
the development of the human faculty of lan-
guage and speech, one can imagine unprece-
dented lines of future research. Comparisons
of the gene in humans to those in chim-
panzees and other primates, and analyses of
the types and patterns of sequence variation
within the region of FOXP2, could add to
our understanding of how human language
evolved
12,13
. An examination of the functions
and expression patterns of the gene (and of
other genes it might set off) in fetal and adult
brain tissue could shed light on how parts of
the human brain are prepared for their role
in cognitive information processing.
The discovery of a gene implicated in
speech and language is among the first fruits
of the Human Genome Project for the cog-
nitive sciences. Just as the 1990s are remem-
bered as the decade of the brain and the
dawn of cognitive neuroscience, the first
decade of the twenty-first century may well
be thought of as the decade of the gene and
the dawn of cognitive genetics.
■
Steven Pinker is in the Department of Brain and
Cognitive Sciences, Massachusetts Institute of
Technology, Cambridge, Massachusetts 02139, USA.
e-mail: steve@psyche.mit.edu
1. Darwin, C. Descent of Man (John Murray, London, 1871).
2. Lai, C. S. L., Fisher, S. E., Hurst, J. A., Vargha-Khadem, F. &
Monaco, A. P. Nature 413, 519–523 (2001).
3. Chomsky, N. Language 35, 26–58 (1959).
4. Lenneberg, E. in The Structure of Language: Readings in the
Philosophy of Language (eds Fodor, J. A. & Katz, J. J.) 579–603
(Prentice-Hall, Englewood Cliffs, NJ, 1964).
5. Stromswold, K. Language (in the press).
6. Hurst, J. A., Baraitser, M., Auger, E., Graham, F. & Norrell, S.
Dev. Med. Child Neurol. 32, 347–355 (1990).
7. Gopnik, M. & Crago, M. Cognition 39, 1–50 (1991).
8. Vargha-Khadem, F. et al. Proc. Natl Acad. Sci. USA 92, 930–933
(1995).
9. Pinker, S. Words and Rules: The Ingredients of Language (Basic
Books, New York, 1999).
10.Fisher, S. E., Vargha-Khadem, F., Watkins, K. E., Monaco, A. P.
& Pembrey, M. E. Nature Genet. 18, 168–170 (1998).
11.Ullman, M. T. & Gopnik, M. Appl. Psycholing. 20, 51–117 (1999).
12.Kreitman, M. Annu. Rev. Genom. Hum. Genet. 1, 539–559 (2000).
13.Aquadro, C. in Limits to Knowledge in Evolutionary Biology (eds
Clegg, M. T., Hecht, M. K. & MacIntyre, J.) 135–149 (Plenum,
New York, 1999).
on the verge of science fiction such as the
quantum computer.
Six years ago a new state of matter
2–7
—
the Bose–Einstein condensate (BEC),
named after those who predicted its exis-
tence — was first created in a dilute gas of
atoms. In a BEC, the usual energy distribu-
tion for an ensemble of particles no longer
exists; all particles are forced to acquire the
same energy. Furthermore, this energy is
always the lowest allowed by quantum
theory; it can be close but not equal to zero.
A BEC contains up to ten million atoms,
all at a temperature just above absolute
zero (a few nanokelvin). In such a state, the
macroscopic cloud of atoms has quantum
features, which are distinctly different from
those of the classical world we observe
around us.
Until now, such clouds of ultracold
atoms have only been handled from a dis-
tance. This is mainly because a BEC is so deli-
cate that any contact with other atoms will
destroy it. For this reason, BEC experiments
are performed inside ultrahigh-vacuum
chambers, providing an environment simi-
lar to that found in space. The clouds are
trapped, manipulated and observed in mag-
netic, electric or light fields, which usually
originate from sources outside the chamber,
such as lasers or magnetic coils. The geome-
try of traps produced by these sources is
therefore limited. A source close to the BEC
could provide much tighter and more com-
plex traps, but there were fears that the
ultralow-temperature cloud would not sur-
vive in the presence of higher-temperature
objects.
The achievement of Reichel and col-
leagues
1
in Munich — and the parallel work
by C. Zimmermann’s group in Tübingen
8
—
is to put the source of the trapping fields
inside the ultrahigh-vacuum chamber, a few
tens of micrometres away from the atom
cloud. The experiments solve both of the
news and views
466 NATURE
|
VOL 413
|
4 OCTOBER 2001
|
www.nature.com
A
toms are the building blocks of all
matter. They have a positively charged
nucleus and their outer boundaries
are defined by electron clouds. They remain
electrically neutral, but the number of elec-
trons governs their chemical properties.
Atoms have long been studied and exploited
by mankind. Yet we are just now learning
a whole new way of communicating with
them. On page 498 of this issue, Reichel
and colleagues
1
describe another step on
this journey. Their achievement may result
in new insights into the foundations of
quantum theory, and lead to applications
Bose–Einstein condensates
Mastering the language of atoms
Ron Folman and Jörg Schmiedmayer
Physicists can already make ultracold atoms perform quantum tricks in
sophisticated magnetic and optical traps. But a fast route to trapping
atoms on a microchip opens up new possibilities.
Many of today’s electronic
devices are unthinkable
without miniaturization. By
similarly shrinking elements
used in atom optics, such
as atom traps, guides,
mirrors, beam-splitters and
interferometers, and by
fabricating them using
modern solid-state
techniques (lithography)
stemming from electronics
and optics, physicists hope
to achieve a similar level of
control over atoms as they
have over electrons and
photons. The preparation,
manipulation and
measurement sensitivity
must reach a level at which
quantum effects are
dominant.
Why use atom chips?
First, studying quantum
behaviour requires the
observed system to be
isolated from its
environment because any
interaction would quickly
destroy the delicate
quantum effects. The neutral
atom is an excellent choice
in this matter — because it
has no charge, it interacts
with its environment in a
relatively weak way.
Second, chips offer a
platform that is robust,
scaleable (it allows for
arrays of traps, for example)
and accurate. Together,
atoms and chips make a
powerful combination.
Lithographic techniques can
now create structures with
length scales below 100
nm, which is smaller than
the quantum-mechanical
(de Broglie) wavelength of
the cooled atoms, ensuring
control at the quantum
level. The small size of the
traps allows atoms to be
positioned in individual
sites separated by small
distances, enabling them
to interact in a controlled
way. Because the atoms
themselves are well
localized (within 10 nm) they
can be manipulated and
detected by miniaturized
light elements, such as
micro-cavities and solid-
state wave guides, which
today can be fabricated on
the same chip.
A long-term goal is to
fabricate everything on the
same chip — from the light
sources (micro-lasers) to
the readout electronics —
producing a truly integrated
self-sufficient device. The
hope is that such devices
will do for quantum atom
optics what integrated
circuits did for electronics.
Atom chips are already an
outstanding research tool.
Perhaps the day is not far
off when they will also be
household items, in clocks,
communications and even
computing. R. F. & J. S.
Box 1 The atom-chip toolbox
© 2001 Macmillan Magazines Ltd